Why momentum is relevant in electoral polling

Despite a chorus of dissent for the concept of electoral momentum, I think the idea has meaning. Put differently, if Mitt Romney has a figurative wind at his back, what is its literal incarnation?

I position electoral momentum within the news cycle, which I have said is neither "24 hour" nor instantaneous. One news cycle is the focus by news media on a single topic/thread/meme (t/t/m) until a related t/t/m emerges with the potency to displace the former -- this is an inflection point, and it moves polls.

The release around Sept 17 of Mitt Romney's "47%" video was a clear inflection point that was widely felt to have secured re-election for the president. But while there was an unmistakable polling trend in favour of Barack Obama following this, another inflection point occurred on Oct 3 as a result of a discrete event -- the 90 minute Colorado debate. When the cameras in Colorado were shut off, Mitt Romney's prospects began to climb daily for more than a week. This begs the key question: why did a great mass of voters only begin to think more favourably of Mitt Romney on Oct 7, for instance? Why did another mass of voters not feel more favourably toward Romney on Oct 7, but did so on Oct 9?

The answer, I believe, is simple if you view the news cycle as a weekly TV drama. From the voter's vantage point, both the 47% video and the Oct 3 debate (and Romney's Benghazi remarks, for that matter) are dramas not unlike an episode of Grey's Anatomy, or other program. Over each week, these voters cross paths with a selection of their acquaintances once: e.g., taking a kid to a ballet recital, attending an ad hoc meeting at work, having lunch with a colleague. These encounters require a topic of conversation, and -- just as Grey's gossip can fill this role, so too can political developments. Thus, there is a multiplier -- the initial news development recurs in a cascade of conversations across the electorate. These conversations educate those who missed the initial news, affirm views, change views, expand knowledge, etc. From the initial pebble toss there is a ripple over about a week.

The Colorado debate t/t/m multiplier, like all trends, faded away, helped in part by the strong Democratic performances in the vice presidential and remaining two presidential debates. But while it lasted, an enormous number of people's views were changed -- in my opinion -- by trusted acquaintances, while at lunches and outside ballet rehearsals.

Edit -- this post is a focused re-write of my Oct 10 post "The very long news cycle"

The very long news cycle

Mitt Romney's strong performance in the first presidential debate this year did not, as of today, cause a spike in his poll results, but rather a steady increase. Which begs the question: if 40 million people watched the 90 minute debate, why would some voters wait a week to change their mind?

I think the answer is in the dynamic of the news cycle.

People use the term "24 hour news cycle", but I think a better analogy is the weekly TV drama. If you're in a part of the population that likes Grey's Anatomy, for example, then perhaps your friends are as well. And after each weekly episode -- and particularly after an emotionally riveting episode -- you don't just turn off the experience at the end of the hour. You talk about it at work the next day. And if your kid has ballet on Thursdays, you talk about it with the adults there. And with the other friends you encounter on the weekend. For the Grey's Anatomy segment, the ripple lasts precisely a week -- until the next episode, at which time interest in the previous episode dies.

Similarly, Mitt Romney's debate performance experienced a multiplier over the last week, likely through millions of conversations as people who encounter other politically-engaged acquaintances need a topic for their conversation.

The same occurred with the 47% video, which had a negative multiplier effect for Mitt Romney, as did his Libya remarks, which were only interrupted by the 47% video. But if there had been an unrelated interruption to the news cycle -- in the example above, perhaps a Grey's Anatomy actor got married -- the previous story would die.

Hours before the VP debate, the game is different. If Biden wins conclusively, the "strong Romney debate / weak Obama debate" story will die and Romney's numbers, I think, will stop moving up. If Ryan wins or the result is inconclusive, the debate story remains ascendant and Romney's trajectory continues. And if a major political event occurs, on the level of the Benghazi attack, the debate story will die.